Modelpublic

worthdoing/Qwen2.5-Coder-7B-Instruct-GGUF

sourceHugging Faceapache-2.0updated 5mo agoView on Hugging Face

0likes152downloads

Model Card

<img src="https://raw.githubusercontent.com/Worth-Doing/brand-assets/main/png/variants/04-horizontal.png" alt="worthdoing" width="400"/> Author: Simon-Pierre Boucher

Qwen2.5-Coder-7B-Instruct - GGUF Quantized by worthdoing

Quantized for local Mac inference (Apple Silicon / Metal) by worthdoing

About

This is a GGUF quantized version of Qwen2.5-Coder-7B-Instruct, optimized for running locally on Apple Silicon Macs with llama.cpp, Ollama, or LM Studio.

—Original model: Qwen/Qwen2.5-Coder-7B-Instruct
—Parameters: 7B
—Quantized by: worthdoing
—Pipeline: corelm-model v1.0

Description

Qwen's dedicated coding model. Top-tier code generation and understanding.

Available Quantizations

File	Quant	BPW	Size	Use Case
`qwen2.5-coder-7b-instruct-Q4_K_M-worthdoing.gguf`	Q4KM	4.58	~3.7 GB	Recommended - Best quality/size ratio
`qwen2.5-coder-7b-instruct-Q5_K_M-worthdoing.gguf`	Q5KM	5.33	~4.3 GB	Higher quality, still fast
`qwen2.5-coder-7b-instruct-Q8_0-worthdoing.gguf`	Q8_0	7.96	~6.5 GB	Near-original quality

How to Use

With Ollama

bash

# Create a Modelfile
cat > Modelfile <<'MODELEOF'
FROM ./qwen2.5-coder-7b-instruct-Q4_K_M-worthdoing.gguf
MODELEOF

ollama create qwen2.5-coder-7b-instruct -f Modelfile
ollama run qwen2.5-coder-7b-instruct

With llama.cpp

bash

llama-cli -m qwen2.5-coder-7b-instruct-Q4_K_M-worthdoing.gguf -p "Your prompt here" -ngl 99

With LM Studio

1.Download the GGUF file
2.Open LM Studio -> My Models -> Import
3.Select the GGUF file and start chatting

Quantization Method

Our quantization pipeline (corelm-model v1.0) follows a rigorous multi-step process to ensure maximum quality and compatibility:

Step 1 — Download & Validation

—Model weights are downloaded from HuggingFace Hub in SafeTensors format (.safetensors)
—Legacy formats (.bin, .pt) are excluded to ensure clean, verified weights
—Tokenizer, configuration, and all metadata are preserved

Step 2 — Conversion to GGUF F16 Baseline

—The original model is converted to GGUF format at FP16 precision using convert_hf_to_gguf.py from llama.cpp
—This lossless baseline preserves the full original model quality
—Architecture-specific tensors (attention, FFN, embeddings, MoE routing) are mapped to their GGUF equivalents

Step 3 — K-Quant Quantization

—The F16 baseline is quantized using llama-quantize with k-quant methods
—K-quants use a mixed-precision approach: more important layers (attention, output) retain higher precision, while less sensitive layers (FFN) are compressed more aggressively
—Each quantization level offers a different quality/size tradeoff:

Method	Bits per Weight	Strategy
Q4_K_M	~4.58 bpw	Mixed 4/5-bit. Attention & output layers use Q5K, FFN layers use Q4K. Best balance of quality and size.
Q5_K_M	~5.33 bpw	Mixed 5/6-bit. Attention & output layers use Q6K, FFN layers use Q5K. Higher quality with moderate size increase.
Q8_0	~7.96 bpw	Uniform 8-bit. All layers quantized to 8-bit. Near-lossless quality, largest file size.

Step 4 — Metadata Injection

—Custom metadata is embedded directly in each GGUF file:
—general.quantized_by: worthdoing
—general.quantization_version: corelm-1.0
—This ensures full traceability and provenance of every quantized file

Tools & Environment

—llama.cpp: Used for both conversion and quantization — the industry-standard open-source LLM inference engine
—Target platform: Apple Silicon Macs (M1/M2/M3/M4) with Metal GPU acceleration
—Inference runtimes: Compatible with llama.cpp, Ollama, LM Studio, koboldcpp, and any GGUF-compatible runtime

Recommended Hardware

Quant	Min RAM	Recommended
Q4KM	4 GB	Mac with 8 GB+ RAM
Q5KM	5 GB	Mac with 8 GB+ RAM
Q8_0	8 GB	Mac with 12 GB+ RAM